PERANCANGAN SISTEM IDENTIFIKASI PENYAKIT UDANG VANNAMEI (LITOPENAEUS VANNAMEI) PADA BUDIDAYA TAMBAK BERBASIS PROGRESSIVE WEB APP DENGAN PEMANFAATAN TEKNOLOGI DEEP LEARNING

Dewi, Nadila (2026) PERANCANGAN SISTEM IDENTIFIKASI PENYAKIT UDANG VANNAMEI (LITOPENAEUS VANNAMEI) PADA BUDIDAYA TAMBAK BERBASIS PROGRESSIVE WEB APP DENGAN PEMANFAATAN TEKNOLOGI DEEP LEARNING. Diploma thesis, UIN RADEN INTAN LAMPUNG.

	PDF Download (11MB)
	PDF Restricted to Repository staff only Download (19MB)

Abstract

ABSTRAK Budidaya udang tambak merupakan salah satu usaha yang menempati posisi strategis dalam sektor perikanan dengan memiliki nilai ekonomi tinggi di Indonesia. Udang Vannamei (Litopenaeus Vannamei) adalah jenis udang yang masuk ke dalam famili Penaeidae, yaitu kelompok udang penaeid yang umumnya hidup di perairan laut dan payau. Produktivitas Udang Vannamei kerap mengalami penurunan akibat serangan penyakit. Jenis penyakit yang umum ditemukan pada udang jenis ini seperti Black Gill dan White Spot Syndrome Virus. Proses identifikasi penyakit udang selama ini masih dilakukan secara konvensional cenderung bersifat subjektif sehingga berpotensi menimbulkan kesalahan dalam diagnosis akibat perbedaan persepsi visual antar manusia. Solusi yang ditawarkan berupa pemanfaatan teknologi Computer Vision dan Deep Learning dalam proses identifikasi penyakit udang Vannamei berdasarkan citra yang ditangkap oleh kamera. Penelitian ini bertujuan untuk merancang sistem identifikasi penyakit udang Vannamei berbasis Progressive Web App guna membantu proses deteksi dini penyakit secara cepat dan akurat pada budidaya tambak udang di Kabupaten Kaur, Provinsi Bengkulu. Metode penelitian yang digunakan adalah Waterfall sebagai metode pengembangan sistem. Deep Learning berupa arsitektur Convolutional Neural Network (CNN) dan model pra-latih YOLOv11 sebagai kerangka pengembangan Object Detection. Dataset yang digunakan dalam penelitian ini sebanyak 3.301 citra telah diaugmentasi terbagi atas 70:10:20 yaitu 70% train (2.312 gambar), 20% validation (659 gambar), 10% test (330 gambar) dan dilatih menggunakan epoch 200. Pengolahan dataset, augmentasi citra, hingga pelatihan model dilakukan menggunakan platform Roboflow dan Google Colab guna mendukung proses komputasi berbasis cloud. Hasil penelitian menunjukkan bahwa model yang dirancang menghasilkan nilai precision sebesar 0,985, recall sebesar 0,977, mean Average Precision (mAP50) sebesar 0,992, rata-rata overlap pada setiap IoU (mAP50-95) sebesar 0,850 serta F1-Score sebesar 0,955. Tingkat akurasi mencapai 94,7% dengan rata-rata waktu pemrosesan identifikasi sebesar 2,7 ms per gambar. Hasil pengujian Black Box Testing menunjukkan bahwa seluruh fitur pada sistem dapat berjalan dengan baik sesuai fungsinya. Berdasarkan hasil tersebut, sistem identifikasi penyakit udang Vannamei dengan memanfaatkan Convolutional Neural Network (CNN) dan model pra-latih YOLOv11 berbasis Progressive Web App mampu mengidentifikasi penyakit udang Vannamei berupa Black Gill dan White Spot Syndrome Virus dengan tingkat akurasi yang sangat baik. Kata Kunci: Deep Learning; Penyakit Udang Vannamei. ABSTRACT Shrimp aquaculture is one of the strategic sectors in the fisheries industry and has high economic value in Indonesia. Vannamei shrimp (Litopenaeus vannamei) belongs to the family Penaeidae, a group of penaeid shrimp commonly found in marine and brackish waters. However, the productivity of Vannamei shrimp frequently declines due to disease outbreaks. Common diseases affecting this species include Black Gill and White Spot Syndrome Virus. Disease identification has traditionally been carried out using conventional methods, which tend to be subjective and may lead to diagnostic errors due to differences in human visual perception. To address this issue, Computer Vision and Deep Learning technologies are proposed for the identification of Vannamei shrimp diseases based on images captured by cameras. Therefore, this study aims to develop a Progressive Web App based Vannamei shrimp disease identification system to support rapid and accurate early disease detection in shrimp farming activities in Kaur Regency, Bengkulu Province. The research employed the Waterfall model as the system development methodology. The Deep Learning approach utilized a Convolutional Neural Network (CNN) architecture and a pre-trained YOLOv11 model as the Object Detection framework. The dataset consisted of 3,301 augmented images, which were divided into 70% training data (2,312 images), 20% validation data (659 images), and 10% Testing data (330 images), and trained for 200 epochs. Dataset preprocessing, image augmentation, and model training were conducted using Roboflow and Google Colab platforms to support cloud based computation. The results demonstrated that the proposed model achieved a precision score of 0.985, recall of 0.977, mean Average Precision (mAP50) of 0.992, mean Average Precision at IoU thresholds ranging from 0.50 to 0.95 (mAP50-95) of 0.850, and an F1-score of 0.955. The system achieved an overall accuracy of 94.7% with an average identification processing time of 2.7 ms per image. Furthermore, Black Box Testing results indicated that all system features functioned properly as intended. Based on these findings, the Progressive Web App based Vannamei shrimp disease identification system utilizing a Convolutional. Neural Network (CNN) and a pre-trained YOLOv11 model was able to accurately identify Black Gill and White Spot Syndrome Virus diseases with excellent performance. Keywords: Deep Learning; Vannamei Shrimp Disease.

Actions (login required)

View Item